One of my main study goals for 2026 is to gain a deeper understanding of Machine Learning. It was previously a sub-discipline of AI, but has become synonymous with it — the impetus behind the latest breakthroughs like ChatGPT.
Section I
Training Setup
You actually don't need a powerful laptop. I'm doing fine thus far on an ancient 6th-gen Intel i7, and when I hit the more compute-intensive stages I'll sign up for Google Colab to leverage cloud compute.
Hardware
- → HP ab292nr — Intel i7-6700HQ · 16 GB RAM · 1 TB SSD
- → Google Colab — Cloud compute for heavier model training
Software Stack
- → VS Code — Primary code editor / IDE
- → Python — scikit-learn, numpy, pandas, matplotlib, torch, streamlit
- → GitHub — Version control
- → Docker — Containerized model serving
- → Streamlit.io — Lightweight app deployment for experiments
Vibe Coding Assistants
- → ChatGPT
- → Claude
Section II
Claude & Vibe Coding
I'm leveraging Claude for vibe coding and looking up definitions, terms, and validating concepts from my training. It is also one of the first things most people consider when they think of AI — a useful reference point for understanding the field itself. I also tried ChatGPT, but have found Claude is better.
Section III
Online ML Courses — Introduction
Starting with the Google Machine Learning Crash Course as my December 2025 study focus, followed by a more structured MOOC. Learning terminology is especially important — knowing that "stochastic" means random or that a "token" represents a word makes abstract concepts far easier to grasp.
- → Google — Machine Learning Crash Course
- → Google — Machine Learning Glossary
Increasing study time from 1–2 hours per week to 30–40 hours per week — roughly 5 hours a day. By January 2026, the goal is to complete the Google course, internalize the glossary, and clearly understand the core concepts.
Section IV
Fundamentals — Math
Next would be brushing up on Math. It's been a while since Calculus in college — hoping it's not too difficult to pick up again. Considering the 12-week Coursera Math for ML and DS specialization.
- → Coursera — Mathematics for Machine Learning and Data Science Specialization
Section V
Statistics
There is an edX Harvard course on Statistics 110: Probability. I'll have to decide which MOOC to standardize on — I don't want to spend money on both Coursera and edX. Might try the free book and lectures first.
- → Harvard / edX — Statistics 110: Probability
Section VI
Python ML Projects
Along the way, experimenting in Python — starting with simple ML projects like email spam detection. The focus is on building strong fundamentals: knowledge that will remain valuable even as tools and trends evolve.
- → O'Reilly — Hands On Machine Learning with Scikit-Learn, Keras & PyTorch
Starting with Scikit-Learn, progressing to PyTorch. Building working models — not just reading about them.
Section VII
MLOps + ML Systems
Need to learn more about Machine Learning Ops — model serving, monitoring & retraining, feature pipelines. While not true MLOps at scale, on a budget Docker and Streamlit.io cover the essentials for small experiments.
- → Docker — Containerization and model serving
- → Streamlit.io — Free, lightweight app deployment for small ML experiments
Section VIII
Learning Plan & Schedule
| Phase | Dates | Duration | Focus |
|---|---|---|---|
| Phase 1 | 12/15/25 – 1/21/26 | 6 weeks | Google ML Vocabulary ✓ done 1/11/26 Google ML Crash Course ✓ done 1/21/26 |
| Phase 2 | 1/22/26 – 4/19/26 | 12 weeks | Statistics 110 · HOMLP · Scikit-Learn & PyTorch model builds · Docker App · Streamlit.io App |
| Phase 3 | 4/20/26 – 7/12/26 | 12 weeks | Coursera Math for ML/DS · Advanced Scikit-Learn & PyTorch model builds |
Comments
Post a Comment